Evaluation of Athlete’s Competitive Ability based on XLNet Transformer and Kolmogorov-Arnold Networks

In recent days, sports are organized as a physical activities which include games, competition, and skills which enhance fitness, entertainment, and motivate for success. It also motivates aspects like discipline, and teamwork. Additionally, evaluation of athlete’s competitive ability is complex because of multifaceted skills such as mental resilience. Moreover, standardized assessments are not collecting sport specific types which lead to incomplete evaluation. It is also difficult replicate the performance because it might be influenced because of stress and competition. To overcome these issues paper proposed a model for Evaluation of Athletes’ Competitive Ability XLNet with Kolmogorov-Arnold Networks (KAN) for evaluation, effectively achieving better accuracy. Initially, the questionnaire is generated by using Technology Acceptance Model (TAM 3) then further filtration and evaluation is performed through Confirmatory Factor Analysis (CFA) and Structural Equation Model (SEM). Then, after the evaluation the optimal attributes are trained using XLNet transformer. Later the trained attributes are classified using Kolmogorov-Arnold Networks (KAN) then finally the evaluation is performed using confusion metrics. The performance of the proposed method is evaluated using metrics like accuracy, precision, and recall respectively.